Re: [PATCH v2 3/3] selftests/x86, x86/ldt: Add a selftest for modify_ldt
On Tue, Jul 21, 2015 at 4:36 PM, Willy Tarreau wrote: > Hi Andy, > > On Tue, Jul 21, 2015 at 12:59:31PM -0700, Andy Lutomirski wrote: >> This tests general modify_ldt behavior (only writes, so far) as >> well as synchronous updates via IPI. It fails on old kernels. >> >> I called this ldt_gdt because I'll add set_thread_area tests to >> it at some point. > > Quick feedback : at two places you have this : > >> + } else if (errno == ENOSYS) { >> + printf("[OK]\tmodify_ldt is returned -ENOSYS\n"); > > => s/is // > > Please add stdlib.h to avoid this warning I'm getting on 32-bit : > > ldt_gdt.c:286:4: warning: incompatible implicit declaration of built-in > function 'exit' [enabled by default] > > And I had to remove xmmintrinsic as suggested by Boris as well. > > > FWIW here's what I'm getting here on 4.1.2 without CONFIG_X86_16BIT > (where nmi_espfix failed), the output is the same for a 32- and a 64-bit > process : > > [OK]LDT entry 0 has AR 0x0040FA00 and limit 0x000A > [OK]LDT entry 0 has AR 0x00C0FA00 and limit 0xAFFF > [OK]LDT entry 1 is invalid > [OK]LDT entry 2 has AR 0x00C0FA00 and limit 0xAFFF > [OK]LDT entry 1 is invalid > [OK]LDT entry 2 has AR 0x00C0FA00 and limit 0xAFFF > [OK]LDT entry 2 has AR 0x00D0FA00 and limit 0xAFFF > [OK]LDT entry 2 has AR 0x00D07A00 and limit 0xAFFF > [OK]modify_ldt rejected 16 bit segment > [OK]LDT entry 2 has AR 0x00D07200 and limit 0xAFFF > [OK]LDT entry 2 has AR 0x00D07000 and limit 0xAFFF > [OK]LDT entry 2 has AR 0x00D07400 and limit 0xAFFF > [OK]LDT entry 2 has AR 0x00507600 and limit 0x000A > [OK]LDT entry 2 has AR 0x00507E00 and limit 0x000A > [OK]LDT entry 2 has AR 0x00507C00 and limit 0x000A > [OK]LDT entry 2 has AR 0x00507A00 and limit 0x000A > [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A > [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A > [RUN] Test fork > [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A > [OK]LDT entry 1 is invalid > [OK]Child succeeded > [OK]modify_ldt failure 22 > [OK]modify_ldt rejected 16 bit segment > [OK]modify_ldt rejected 16 bit segment > [OK]modify_ldt rejected 16 bit segment > [OK]modify_ldt rejected 16 bit segment > [OK]modify_ldt rejected 16 bit segment > [OK]modify_ldt rejected 16 bit segment > [OK]LDT entry 0 is invalid > [OK]LDT entry 0 has AR 0x0040F200 and limit 0x > [OK]LDT entry 0 is invalid > [RUN] Cross-CPU LDT invalidation > [FAIL] 5 of 5 iterations failed That's intentional. Old modify_ldt can leave other threads with stale cached descriptors, which is what you're seeing. --Andy > > Thanks, > Willy > -- Andy Lutomirski AMA Capital Management, LLC -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 3/3] selftests/x86, x86/ldt: Add a selftest for modify_ldt
Hi Andy, On Tue, Jul 21, 2015 at 12:59:31PM -0700, Andy Lutomirski wrote: > This tests general modify_ldt behavior (only writes, so far) as > well as synchronous updates via IPI. It fails on old kernels. > > I called this ldt_gdt because I'll add set_thread_area tests to > it at some point. Quick feedback : at two places you have this : > + } else if (errno == ENOSYS) { > + printf("[OK]\tmodify_ldt is returned -ENOSYS\n"); => s/is // Please add stdlib.h to avoid this warning I'm getting on 32-bit : ldt_gdt.c:286:4: warning: incompatible implicit declaration of built-in function 'exit' [enabled by default] And I had to remove xmmintrinsic as suggested by Boris as well. FWIW here's what I'm getting here on 4.1.2 without CONFIG_X86_16BIT (where nmi_espfix failed), the output is the same for a 32- and a 64-bit process : [OK]LDT entry 0 has AR 0x0040FA00 and limit 0x000A [OK]LDT entry 0 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 1 is invalid [OK]LDT entry 2 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 1 is invalid [OK]LDT entry 2 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D0FA00 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07A00 and limit 0xAFFF [OK]modify_ldt rejected 16 bit segment [OK]LDT entry 2 has AR 0x00D07200 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07000 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07400 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00507600 and limit 0x000A [OK]LDT entry 2 has AR 0x00507E00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507C00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507A00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [RUN] Test fork [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [OK]LDT entry 1 is invalid [OK]Child succeeded [OK]modify_ldt failure 22 [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]LDT entry 0 is invalid [OK]LDT entry 0 has AR 0x0040F200 and limit 0x [OK]LDT entry 0 is invalid [RUN] Cross-CPU LDT invalidation [FAIL] 5 of 5 iterations failed Thanks, Willy -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 3/3] selftests/x86, x86/ldt: Add a selftest for modify_ldt
On Tue, Jul 21, 2015 at 3:02 PM, Boris Ostrovsky wrote: > On 07/21/2015 03:59 PM, Andy Lutomirski wrote: >> >> diff --git a/tools/testing/selftests/x86/ldt_gdt.c >> b/tools/testing/selftests/x86/ldt_gdt.c >> new file mode 100644 >> index ..6f6699f0351a >> --- /dev/null >> +++ b/tools/testing/selftests/x86/ldt_gdt.c >> @@ -0,0 +1,492 @@ >> +/* >> + * ldt_gdt.c - Test cases for LDT and GDT access >> + * Copyright (c) 2015 Andrew Lutomirski >> + */ >> + >> +#define _GNU_SOURCE >> +#include >> +#include >> +#include >> +#include >> +#include >> +#include >> +#include >> +#include > > > Is xmmintrin.h necessary? It breaks 32-bit build with no, and I'll remove it. --Andy -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 3/3] selftests/x86, x86/ldt: Add a selftest for modify_ldt
On 07/21/2015 03:59 PM, Andy Lutomirski wrote: diff --git a/tools/testing/selftests/x86/ldt_gdt.c b/tools/testing/selftests/x86/ldt_gdt.c new file mode 100644 index ..6f6699f0351a --- /dev/null +++ b/tools/testing/selftests/x86/ldt_gdt.c @@ -0,0 +1,492 @@ +/* + * ldt_gdt.c - Test cases for LDT and GDT access + * Copyright (c) 2015 Andrew Lutomirski + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include Is xmmintrin.h necessary? It breaks 32-bit build with /usr/lib/gcc/x86_64-redhat-linux/4.7.2/include/xmmintrin.h:32:3: error: #error "SSE instruction set not enabled" unless I add -msse2. (This also needs stdlib.h for exit() declaration) -boris -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 3/3] selftests/x86, x86/ldt: Add a selftest for modify_ldt
On 07/21/2015 03:59 PM, Andy Lutomirski wrote: diff --git a/tools/testing/selftests/x86/ldt_gdt.c b/tools/testing/selftests/x86/ldt_gdt.c new file mode 100644 index ..6f6699f0351a --- /dev/null +++ b/tools/testing/selftests/x86/ldt_gdt.c @@ -0,0 +1,492 @@ +/* + * ldt_gdt.c - Test cases for LDT and GDT access + * Copyright (c) 2015 Andrew Lutomirski + */ + +#define _GNU_SOURCE +#include err.h +#include stdio.h +#include stdint.h +#include signal.h +#include setjmp.h +#include string.h +#include errno.h +#include xmmintrin.h Is xmmintrin.h necessary? It breaks 32-bit build with /usr/lib/gcc/x86_64-redhat-linux/4.7.2/include/xmmintrin.h:32:3: error: #error SSE instruction set not enabled unless I add -msse2. (This also needs stdlib.h for exit() declaration) -boris -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 3/3] selftests/x86, x86/ldt: Add a selftest for modify_ldt
On Tue, Jul 21, 2015 at 3:02 PM, Boris Ostrovsky boris.ostrov...@oracle.com wrote: On 07/21/2015 03:59 PM, Andy Lutomirski wrote: diff --git a/tools/testing/selftests/x86/ldt_gdt.c b/tools/testing/selftests/x86/ldt_gdt.c new file mode 100644 index ..6f6699f0351a --- /dev/null +++ b/tools/testing/selftests/x86/ldt_gdt.c @@ -0,0 +1,492 @@ +/* + * ldt_gdt.c - Test cases for LDT and GDT access + * Copyright (c) 2015 Andrew Lutomirski + */ + +#define _GNU_SOURCE +#include err.h +#include stdio.h +#include stdint.h +#include signal.h +#include setjmp.h +#include string.h +#include errno.h +#include xmmintrin.h Is xmmintrin.h necessary? It breaks 32-bit build with no, and I'll remove it. --Andy -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 3/3] selftests/x86, x86/ldt: Add a selftest for modify_ldt
Hi Andy, On Tue, Jul 21, 2015 at 12:59:31PM -0700, Andy Lutomirski wrote: This tests general modify_ldt behavior (only writes, so far) as well as synchronous updates via IPI. It fails on old kernels. I called this ldt_gdt because I'll add set_thread_area tests to it at some point. Quick feedback : at two places you have this : + } else if (errno == ENOSYS) { + printf([OK]\tmodify_ldt is returned -ENOSYS\n); = s/is // Please add stdlib.h to avoid this warning I'm getting on 32-bit : ldt_gdt.c:286:4: warning: incompatible implicit declaration of built-in function 'exit' [enabled by default] And I had to remove xmmintrinsic as suggested by Boris as well. FWIW here's what I'm getting here on 4.1.2 without CONFIG_X86_16BIT (where nmi_espfix failed), the output is the same for a 32- and a 64-bit process : [OK]LDT entry 0 has AR 0x0040FA00 and limit 0x000A [OK]LDT entry 0 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 1 is invalid [OK]LDT entry 2 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 1 is invalid [OK]LDT entry 2 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D0FA00 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07A00 and limit 0xAFFF [OK]modify_ldt rejected 16 bit segment [OK]LDT entry 2 has AR 0x00D07200 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07000 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07400 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00507600 and limit 0x000A [OK]LDT entry 2 has AR 0x00507E00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507C00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507A00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [RUN] Test fork [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [OK]LDT entry 1 is invalid [OK]Child succeeded [OK]modify_ldt failure 22 [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]LDT entry 0 is invalid [OK]LDT entry 0 has AR 0x0040F200 and limit 0x [OK]LDT entry 0 is invalid [RUN] Cross-CPU LDT invalidation [FAIL] 5 of 5 iterations failed Thanks, Willy -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 3/3] selftests/x86, x86/ldt: Add a selftest for modify_ldt
On Tue, Jul 21, 2015 at 4:36 PM, Willy Tarreau w...@1wt.eu wrote: Hi Andy, On Tue, Jul 21, 2015 at 12:59:31PM -0700, Andy Lutomirski wrote: This tests general modify_ldt behavior (only writes, so far) as well as synchronous updates via IPI. It fails on old kernels. I called this ldt_gdt because I'll add set_thread_area tests to it at some point. Quick feedback : at two places you have this : + } else if (errno == ENOSYS) { + printf([OK]\tmodify_ldt is returned -ENOSYS\n); = s/is // Please add stdlib.h to avoid this warning I'm getting on 32-bit : ldt_gdt.c:286:4: warning: incompatible implicit declaration of built-in function 'exit' [enabled by default] And I had to remove xmmintrinsic as suggested by Boris as well. FWIW here's what I'm getting here on 4.1.2 without CONFIG_X86_16BIT (where nmi_espfix failed), the output is the same for a 32- and a 64-bit process : [OK]LDT entry 0 has AR 0x0040FA00 and limit 0x000A [OK]LDT entry 0 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 1 is invalid [OK]LDT entry 2 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 1 is invalid [OK]LDT entry 2 has AR 0x00C0FA00 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D0FA00 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07A00 and limit 0xAFFF [OK]modify_ldt rejected 16 bit segment [OK]LDT entry 2 has AR 0x00D07200 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07000 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00D07400 and limit 0xAFFF [OK]LDT entry 2 has AR 0x00507600 and limit 0x000A [OK]LDT entry 2 has AR 0x00507E00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507C00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507A00 and limit 0x000A [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [RUN] Test fork [OK]LDT entry 2 has AR 0x00507800 and limit 0x000A [OK]LDT entry 1 is invalid [OK]Child succeeded [OK]modify_ldt failure 22 [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]modify_ldt rejected 16 bit segment [OK]LDT entry 0 is invalid [OK]LDT entry 0 has AR 0x0040F200 and limit 0x [OK]LDT entry 0 is invalid [RUN] Cross-CPU LDT invalidation [FAIL] 5 of 5 iterations failed That's intentional. Old modify_ldt can leave other threads with stale cached descriptors, which is what you're seeing. --Andy Thanks, Willy -- Andy Lutomirski AMA Capital Management, LLC -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/